Claude Code
Anthropic's terminal-based agentic coding assistant that edits code and runs commands.
Liked as a smooth daily coding agent123 that works well for code review and human-in-the-loop development4567, and some prefer it in the terminal over Cursor.89 The most common complaint is the quality of the harness itself, with commenters calling it buggy, getting worse, or sloppily built10111213, alongside worries about token burn and usage limits.1415
Why people recommend it
Daily reliability · 3
Heavy daily users say it works smoothly and rarely hits bugs.
“I use Claude Code every day. And I find the product to work smoothly, and effectively.”
“Claude code works well enough. I don't encounter many bugs, and I use it for 6-12 hours a day.”
Show 1 more
“I use Claude Code every day. And it is good. Ahead of the pack.”
Code review · 3
Commenters use it to review codebases and PRs and say it catches more than human reviewers.
“I use Claude Code for code review, and honestly it catches more than human reviewers have in my experience.”
“ask Claude Code with Opus 5.5 to review your codebase or a PR or two”
Show 1 more
“I generally start a new Claude Code and a Codex and have them review the code”
Collaborative workflow · 3
Works well with back-and-forth discussion, guidance and review rather than unattended runs.
“I just use Claude Code with a good bit of back-and-forth and review.”
“I have much better real world output with the Claude Code style human in the loop approach”
Show 1 more
“I work by having a discussion with claude code about what I'm trying to do and how it might be built.”
Beyond coding · 3
Used for personal automation, self-hosting, sysadmin work and reverse engineering instrument data.
“Claude Code/Codex has made my entire macOS environment as malleable as Emacs. That’s huge!”
“Codex/Claude Code alongside things like Cloudflare and Tailscale make self hosting far easier”
Show 1 more
“claude code helped me get access to the raw files from a scientific instrument”
Terminal polish · 3
Described as polished, faster and smoother in the terminal, enough to replace Cursor for some.
“I still stick to Claude Code in terminal though, it just seems faster and smoother”
Show 1 more
“when I used Claude Code for the first time, I never went back to Cursor.”
Parallel agents · 3
Runs multiple agent sessions at once, with one user never hitting subscription limits.
“Claude Code will allow you to run multiple instances in the same folder and let them communicate.”
“have never hit a limit”
Show 1 more
“run Claude Code agents in parallel for the implementation”
Caveats and dissent
Harness quality · 3
Some call the harness bad, worsening, or sloppily built, citing unreliable sub-agents and leaked code.
“I was shocked by how bad the harness is and how its only gotten worse since spring?”
“Stay away from OpenCode, its a pile of slop like Claude Code.”
Show 1 more
“nobody there reads their pushed code anymore. Take a look at the previous Claude Code leaked codebases”
Tokens and limits · 2
Newer versions reportedly burn more tokens, and the 5-hour limit can block other Claude use.
“The newer the CC cli is, the more tokens it burns.”
“I cant ask a simple question to Claude, not even to Haiku, because I used my claude code 5h limit.”
Unwanted behaviors · 3
Commenters object to forced auto-updates, resetting the effort setting, and other behaviors patched out by packagers.
“Claude Code will refuse to remember a setting of max and will continually reset it to xhigh”
“As a result, I never knew that Claude Code does a nasty thing like this, either.”
Show 1 more
“The newer the CC cli is, the more tokens it burns.”
Email exposure · 1
One commenter says the user's email is placed in the system prompt and sent in requests.
“Anthropic includes your email in the system prompt”
Third-party harnesses · 1
Discrepancies were reported when using it through Zed's External Agent feature.
“I noticed discrepancies when using Zed's External Agent feature with Claude Code”
Often compared with
3 moved from Codex to Claude Code
43 moved from Claude Code to CodexOn the home page's Debates
2 moved from OpenCode to Claude Code
24 moved from Claude Code to OpenCode
22 moved from Cursor to Claude Code
3 moved from Claude Code to CursorPi123 comments mention both
1 moved from Pi to Claude Code
23 moved from Claude Code to Pi
8 moved from GitHub Copilot to Claude Code
3 moved from Claude Code to GitHub Copilot
1 moved from Claude Code to Claude
Who recommends it in their own words
From the makers not counted in the ranking
“It's already been fixed as part of v2.1.281 releasing today.”
“Hey all, Boris from the Claude Code team here.”
“way better than Claude Code with Opus 4.5”
“we shipped our differential renderer to everyone today”
“Hey HN, Boris from the Claude Code team here.”
“We love python, react, tailwind, OpenAI, Google Cloud, and Claude Code. The ideal candidate is running 10-15 instances of Claude Code at once.”
Threads where it came up
- Ask HN: What is your (AI) dev tech stack / workflow? (June 2026) 24 mentions
- Ask HN: Who wants to be hired? (October 2026) 17 mentions
- Auto mode is now the default in Claude Code 14 mentions
- Ask HN: How is AI-assisted coding going for you professionally? 14 mentions
- Ask HN: Is anyone experimenting with different ways of using LLMs for coding? 11 mentions
- Cursor 3 11 mentions
Also written as Claude Code auto mode, Claude Code Auto Mode. Data: claude-code.json. Quotes are verbatim; the verdict and theme labels are written by a model from these comments, and every clause links to its sources.
All comments 47
“Where Claude Code and Codex have made progress recently, I feel, is with sharpening the interactivity.”
“I used Cursor first, tried Claude Code, but went back to Cursor.”
“I use Claude Code, Codex, opencode and a couple other but I do not like the DX/UX.”
“when I used Claude Code for the first time, I never went back to Cursor.”
“My Codex and Claude Code bots (and Copilot at work) run on it pretty much around the clock”
“it runs Claude code, codex, opencode, pi, whatever model is appropriate”
“I use Claude Code everyday in my project, claude code wrote codes and I reviewed them.”
“nobody there reads their pushed code anymore. Take a look at the previous Claude Code leaked codebases”
“I use Claude Code extensively”
“Code is tasked to output a detailed view of what it found, run measurements and tests”
“I use Claude Code extensively.”
“installed claude code and asked it to setup herdr”
“I work by having a discussion with claude code about what I'm trying to do and how it might be built.”
“I still stick to Claude Code in terminal though, it just seems faster and smoother”
“vim, ssh, git, scripts for building, and now Claude code”
“The newer the CC cli is, the more tokens it burns.”
“Stay away from OpenCode, its a pile of slop like Claude Code.”
“I use claude code as my harness”
“Anthropic includes your email in the system prompt”
“I used a prompt similar to what I use with Codex and Claude Code”
“Claude Code would be another example.”
“I just use Claude Code with a good bit of back-and-forth and review.”
“I use that claude plan for 90% of my work, plus a bit of GPT”
“I built it with Claude Code and Codex.”
“I use Claude Code.”
“Claude code is very polished”
“they have a flag on claude code to disable telemetry”
“ask Claude Code with Opus 5.5 to review your codebase or a PR or two”
“I ran the same Python codebase with different levels of type hints through Claude Code and Codex”
“of the set {Claude Code, Codex, Pi, OMP} I must say I find CC to be the worst.”
“I use claude code daily and if the open ai copies it 1$ cheaper per seat my company is switching immediately.”
“Even Claude Code is better than OpenCode and that's saying something.”
“I noticed discrepancies when using Zed's External Agent feature with Claude Code”
“I use Copilot CLI at work, and Claude Code at home, and all harnesses are pretty good these days.”
“I was shocked by how bad the harness is and how its only gotten worse since spring?”
“I'd love to rip out Siri and replace it with Claude Code or something like that.”
“claude code/t3 (with their build it browsers) + grok-bot style agent cover all the bases for me”
“created in 3 weeks using a combination of Claude Code and Codex”
“Now I find the included auto-classifier is very reliable.”
“they drift into the ghastly nightmare of the Anthropocene (go ahead, take a look at the leaked source code for Claude Code)”
“I have much better real world output with the Claude Code style human in the loop approach”
“I use Claude Code every day. And I find the product to work smoothly, and effectively.”
“I use Claude Code for code review, and honestly it catches more than human reviewers have in my experience.”
“I use Claude Code every day. And it is good. Ahead of the pack.”
“Claude code works well enough. I don't encounter many bugs, and I use it for 6-12 hours a day.”
“I use Claude Code and OpenCode almost equally and Codex slightly less and Pi the least of them”
“I cant ask a simple question to Claude, not even to Haiku, because I used my claude code 5h limit.”